jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 23901
Joined: Sat Jul 30, 2011 7:41 pm

Re: Power supply WARNING

Sun Jan 08, 2012 5:18 pm

emilio said:


Can I use a mobile phone charger?


If it pouts out enough current, yes.
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Contrary to popular belief, humorous signatures are allowed. Here's an example...
“I think it’s wrong that only one company makes the game Monopoly.” – Steven Wright

User avatar
Jessie
Posts: 1754
Joined: Fri Nov 04, 2011 7:40 pm
Location: C/S CO USA

Re: Power supply WARNING

Sun Jan 08, 2012 5:46 pm

Cracknel said:


I have seen so many USB power supply units outputting more than 5 volts that I have lost my hope in finding a good one.

Most of them are even labelled with 5.7 volts.



5.7V wouldn't concern me in the slightest, assuming it didn't go over that amount.  The USB  specifacation says 5V +/- 5% which is a max of 5.25.  The size (gauge) of wire that gets used in the average USB cable would cause enough V drop that I'm sure even at 5.7V you would be lucky to get 5V out of the end of the cable.

Although, this thread makes me wonder...  I bought a wall socket with USB ports on it a couple months agao I wonder what it puts out?

User avatar
Jessie
Posts: 1754
Joined: Fri Nov 04, 2011 7:40 pm
Location: C/S CO USA

Re: Power supply WARNING

Sun Jan 08, 2012 9:02 pm

Jongoleur said:


@plugwash:

Just watched the teardown of a USB psu from your link.

Eyewatering.


Something really bugged me about this video and it wasn't until I went to the gym today that it occured to me.  The test that is done in this video with various meeters is nothing more than a horse and pony show.  This is why... Voltage is nothing more than a measure of potential.  According to Ohms law Voltage = Current * Resistance.  So at the point where he shows that voltage has gotten under 1V (around the 4:03 mark) he says something like "that is pretty much a short circuit."  Throughout the video he uses voltage and output in the same sentence.  Current and output go togather.  Anyaway, back to the issue here if you have a short circuit condition you are very close to a 0 ohm resistance cituation (there is no such thing as no resistance.)  So if you take ohms law and I use the figures on his screen (the resistance isn't given, but I could figure it out)  but just to simplify I make resistance 0.. .723 (Amps current) * 0 (Resistance) = 0 Volts.  You see how lowering reistance and puting more load on the power supply lowers the Voltage the DMM will read?  Therefore, that test was a waste of time.  He comments on how melted and hot the thing becomes...  Well yeh he created a direct short and that PSU has pretty much no overcurrent protetection.

I think his assessment of the build quiality is good, but the whole part with the 4 different meeters is a waste of time.  I gaurantee that any power supply out there would exibit the same results the only difference being that most of them would cut off when over-drawn.  If you introduce a short across any type of circuit you will read minimal voltage because voltage is a measure of potential and not a measure of energy flowing.  When you introduce a low resistance short than there is no longer the potential for energy to flow but now there is actual energy flow.

rgh
Posts: 212
Joined: Fri Nov 25, 2011 3:53 pm

Re: Power supply WARNING

Sun Jan 08, 2012 9:23 pm

But it was rated to provide 1A at 5V.  It wouldn't even produce 1A on effective short circuit.  I happen to have one of the same model; if you draw 200mA it drops to about 4V.  It's now in the bin

slacker
Posts: 33
Joined: Wed Dec 07, 2011 9:12 pm

Re: Power supply WARNING

Sun Jan 08, 2012 9:43 pm

Yeah, the test is pointless once he gets past the point where the voltage drops below 5 volts, he's then proved it can t do what it claims.

At the end he s asking it to drive a load of about 1.7 Ohms (1.24 volts/0.725 Amps), which is way beyond what it was ever even claimed to do. 1 Amp at 5 volts being a load of 5 Ohms. Like Jessie said pretty much any power supply would freak out in some way if you asked it to do that.

Tomo2k
Posts: 127
Joined: Mon Dec 19, 2011 10:00 pm

Re: Power supply WARNING

Mon Jan 09, 2012 1:27 am

Erm, no.

Jessie, you haven"t understood.

This device is supposed to be a regulated power supply. Ohms law does not apply.
(Or rather, it should not.)

He is asking the supply to provide a chosen current between 0mA and 1000mA (the nameplate ratings) using a variable-current load.

With a "good" power supply, it would read roughly 5VDC throughout that entire range of current supplied.

All other results are a failure.

As this one is physically incapable of supplying 1000mA, it is an abject failure and should have been reported to Trading Standards as it is clearly not fit for purpose.

Hajj_3: Is this one a clone of the new Apple or Kindle "Type G" "same size as a normal plug" USB PSU?

If so, please discontinue use immediately. At that size it is incredibly difficult to meet the safety requirements. The teardown of one of those fakes I"ve seen had a mains to USB connector clearance of <1mm, and one of the USB-side capacitors was physically pressed against the back of the Neutral pin.

- In other words, shaking it could short the USB to Live, and the USB will short to Neutral after a while, which is equally dangerous.

(To be honest I"m unconvinced the genuine ones are OK, but I"ve not seen inside so I"ll give them the benefit of the doubt.)

I would post your 99p PSU to Trading Standards along with where and how you bought it, saying that you aren"t sure it really meets CE.
– They are starting to take notice and are trying to hold EBay itself to account for selling non-CE power supplies.

There have been more than a few accidents and fires in the last couple of years caused by non-CE compliant PSUs.

Finally, to everyone:

NEVER EVER buy any mains adapters from EBay – USB or otherwise.

They are very likely to be dangerous because if they really could build a USB PSU to CE requirements for 50p then they would be selling them in real shops for 10 quid and making a lot more money.
These people are not in it for fun, they are trying to make money.

If they say the PSU is a "Genuine Apple" or "Genuine Nokia" etc, would you be able to tell if it was a fake or not?

Unfortunately it is getting harder to find good ones – it appears that even some Amazon partners may be trying to pass off fake "Apple" USB chargers occasionally.

For example, this one seems far too cheap, has a lot of very bad reviews including some apparently describing dangerous failures and one reviewer has provided comparison photos which he claims prove it to be a fake.

Rule of thumb: If it"s much less than half the price they are in your local mobile phone shop, then it"s quite likely to be either stolen or fake.

jwatte
Posts: 203
Joined: Sat Aug 13, 2011 7:28 pm

Re: Power supply WARNING

Mon Jan 09, 2012 2:14 am

hippy said:


bf4ed said:


Just add LM7805 before power input of r-pi. For use in car or home its wise to add some voltage stabilizer, just bo be on the safe side. Then you can use a cheap ebay supply.


The problem is that people buying a supply marked 5V and suitable for USB powering will expect it to be a 5V supply with tight regulation to that. They won't know they have to use an additional regulator until possibly too late. Outputting 12V isn't the worse of it; get something really bad and you can find it puts mains out!



Not to mention that the 7805 is a 2V drop-out part -- you have to power it with at least 7V to have to output stable 5V.

I think pretty much any peripheral board for the RPi will need a wall wart input that takes 7V (or more likely 9V -- easier to find).

Also, if the board is already USB powered, meaning it can draw 500 mA, and the board draws a bunch of power, it really can't supply the 500 mA something else will need on the USB port. You'd need a powered hub as the first device out on that bus if you expect to power things like sound cards or tablets or game controllers with force feedback.

User avatar
Chromatix
Posts: 430
Joined: Mon Jan 02, 2012 7:00 pm
Location: Helsinki

Re: Power supply WARNING

Mon Jan 09, 2012 5:43 am

As I understand it, the power input is supposed to be from a phone charger, the "micro USB" socket matching the new EU standard for phone charging.  These can safely provide considerably more than the 500mA available from a standard USB socket.

The "USB wall wart" supplies are also intended for charging phones, and genuine ones likewise tend to provide more power than a standard USB port would.

Some computers can also provide extra power beyond the USB standard, but usually this is made conditional on there actually being a phone at the far end of the cable.  I wouldn't rely on a computer USB port being able to power the R-Pi - although it might work anyway.
The key to knowledge is not to rely on people to teach you it.

User avatar
Jessie
Posts: 1754
Joined: Fri Nov 04, 2011 7:40 pm
Location: C/S CO USA

Re: Power supply WARNING

Mon Jan 09, 2012 5:57 am

Tomo2k said:


Erm, no.

Jessie, you haven"t understood.

This device is supposed to be a regulated power supply. Ohms law does not apply.
(Or rather, it should not.)


I studied in trade school for 4 years and then went through another 4 years as an apprentice electrician.  I also maintain a Colorado Journeyman's Electrical license which I have to re-test for every couple years.  Before that I also recieved an Bachelors of Scinece in IT.  Ohms law always applys or it wouldn't be a law.  It isn't a theroy, it is a law.

I wasn't arguing that these power supplys were good, quite the opposite they are quite poor, however this guys tests are what I question.

Edited: To be less harsh.

User avatar
Chromatix
Posts: 430
Joined: Mon Jan 02, 2012 7:00 pm
Location: Helsinki

Re: Power supply WARNING

Mon Jan 09, 2012 6:15 am

I thought it was quite a good demonstration of the shoddiness of the unit.  First deonstrating that even at a fraction of rated load, it was below nominal voltage, and then showing that it couldn't reach full current even at nil voltage.

Now, Ohm's law does still apply even to a regulated PSU, but the "regulator" part is supposed to automatically compensate, increasing the power in order to keep voltage steady as load resistance drops and current rises.  This PSU clearly did not have a regulator, either by inspection or by results.

Looking that the PSU diagram on the R-Pi, it seems quite sensible to me.  There is obviously a limit to what it can protect against (raw mains...  no) but the combination of the semiconductor fuse and the capacitors should give ample opportunity for the clamp diode to operate in less extreme circumstances, eg. neither a 12V input nor a static discharge should cause damage.
The key to knowledge is not to rely on people to teach you it.

User avatar
Jessie
Posts: 1754
Joined: Fri Nov 04, 2011 7:40 pm
Location: C/S CO USA

Re: Power supply WARNING

Mon Jan 09, 2012 6:23 am

Chromatix said:


Now, Ohm's law does still apply even to a regulated PSU, but the "regulator" part is supposed to automatically compensate, increasing the power in order to keep voltage steady as load resistance drops and current rises.  This PSU clearly did not have a regulator, either by inspection or by results.


You are of course correct, no electrical current is outside of the bounds of Ohms law.  Just that normally these supplies are current or voltage regulated more often the later.  I was not arguing that these supplys were good just that they gave the results expected as they only contained a transformer and no Voltage regulating IC.  I think I was a tad quck to get upset and lash out above, but rather than edit it out (or delete it) I will leave it there to serve as a reminder to my lapse in judgement.

slaeshjag
Posts: 9
Joined: Sat Jan 07, 2012 12:57 pm
Contact: Website

Re: Power supply WARNING

Mon Jan 09, 2012 6:28 am

Jessie said:


Tomo2k said:


Erm, no.

Jessie, you haven"t understood.

This device is supposed to be a regulated power supply. Ohms law does not apply.
(Or rather, it should not.)


I studied in trade school for 4 years and then went through another 4 years as an apprentice electrician.  I also maintain a Colorado Journeyman's Electrical license which I have to re-test for every couple years.  Before that I also recieved an Bachelors of Scinece in IT.  Ohms law always applys or it wouldn't be a law.  It isn't a theroy, it is a law.

I wasn't arguing that these power supplys were good, quite the opposite they are quite poor, however this guys tests are what I question.

Edited: To be less harsh.



Then I would like you to elaborate what it was with the tests that he did wrong, considered this is supposed to be a regulated switched power supply (with your education, that really shouldn't be a problem, right?

The key is he uses a variable load, and he does NOT short circuit it. It is clear though that the powersupply fails to compensate for the load, thereby the voltage drops. This could be seen as a short-circuit situation, but this is a load that the supply is supposed to be able to handle. Therefore, I can't find anything wrong with the method he used to test it, other than the fect that he seemed to have very little protection from sudden component failure, which I was expecting (and when he moved in with his hand to mesure the temperature of the components, I was just waiting for him go get zapped by the mains.)

A constant load is a best-case situation for the supply, as it simulates a current drain without the load spikes that, for example, a seeking (or, a starting) harddisk would produce. And yet it failed.

User avatar
Jessie
Posts: 1754
Joined: Fri Nov 04, 2011 7:40 pm
Location: C/S CO USA

Re: Power supply WARNING

Mon Jan 09, 2012 6:40 am

slaeshjag said:


The key is he uses a variable load, and he does NOT short circuit it. It is clear though that the powersupply fails to compensate for the load, thereby the voltage drops. This could be seen as a short-circuit situation, but this is a load that the supply is supposed to be able to handle. Therefore, I can"t find anything wrong with the method he used to test it, other than the fect that he seemed to have very little protection from sudden component failure, which I was expecting (and when he moved in with his hand to mesure the temperature of the components, I was just waiting for him go get zapped by the mains.)


Yeh he does short the load, infact, if you watch the video he says it at around the 4:03 mark.

No where do I say the power supply is doing what it should only that it doing as expected under Ohm's laws.  The power supply he is useing is clearly not a self regulating unit as proven when he opens up the unit in the video.  There is no Voltage regulating IC on the board.  So therefore you cannont expect it to increase the voltage or current when under load.  His assessment of the poor build of the PSU is correct but showing us that there is little difference in potential is where I had the issue.

slaeshjag
Posts: 9
Joined: Sat Jan 07, 2012 12:57 pm
Contact: Website

Re: Power supply WARNING

Mon Jan 09, 2012 7:01 am

Guess I didn't know my electronics as good as I though, can't find a way for a flyback converter to compensate without feedback from the output side of the transformer.

error404
Posts: 351
Joined: Wed Dec 21, 2011 11:49 pm

Re: Power supply WARNING

Mon Jan 09, 2012 8:09 am

Jessie said:


There is no Voltage regulating IC on the board.  So therefore you cannont expect it to increase the voltage or current when under load.  His assessment of the poor build of the PSU is correct but showing us that there is little difference in potential is where I had the issue.


There isn't a regulator IC, however it is certainly possible to build a constant voltage source without one, but his point is to evaluate its label claims, not whether the (shoddy and insufficient) design behaves as he expects (and which most YT viewers are no doubt incapable of evaluating). He is showing that this is not a constant voltage supply, and that it can't live up to its claimed output specification of 1A @ 5V either. It can't even do 1A at all. I don't see where you take issue with this test, it's perfectly valid and a pretty standard way of load testing a regulated supply.

At least the supply in question never outputs a dangerous voltage and actually has some kind of mains isolation though, I've seen far worse. Seriously be careful with grey market power supplies, they can be very dangerously designed.


Guess I didn't know my electronics as good as I though, can't find a way for a flyback converter to compensate without feedback from the output side of the transformer.


LOL yes. With some cleverness you might be able to make it work by monitoring primary current...

adlambert

Re: Power supply WARNING

Mon Jan 09, 2012 8:23 am

Jessie said:


Tomo2k said:


Erm, no.

Jessie, you haven"t understood.

This device is supposed to be a regulated power supply. Ohms law does not apply.
(Or rather, it should not.)


I studied in trade school for 4 years and then went through another 4 years as an apprentice electrician.  I also maintain a Colorado Journeyman's Electrical license which I have to re-test for every couple years.  Before that I also recieved an Bachelors of Scinece in IT.  Ohms law always applys or it wouldn't be a law.  It isn't a theroy, it is a law.

I wasn't arguing that these power supplys were good, quite the opposite they are quite poor, however this guys tests are what I question.

Edited: To be less harsh.


This was an extreme test to demonstrate the lack of claimed regulation, and with that aim it was perfectly valid and successful.

The purpose of a (voltage) regulator is to maintain constant voltage despite changes in current draw and other variables, so it is supposed to defy ohms law within certain limits, and well designed regulators do that.

If the output was unregulated then I would expect your critisism to apply, but the test proved the inadequate regulation so I am happy with it.

Anyway, not a big deal for most of us reading this, low cost, good power supplies are not hard to come by, I have a drawer full.  Those folk that have chosen R-Pi because they are cash strapped will just need sufficient warning about the really bad products out there.

User avatar
Burngate
Posts: 6059
Joined: Thu Sep 29, 2011 4:34 pm
Location: Berkshire UK Tralfamadore
Contact: Website

Re: Power supply WARNING

Mon Jan 09, 2012 12:36 pm

Just to obfuscate: Ohms law doesn't always apply - think of a zener!

Also, his extra meters were to show that not only is the PSU horrible, but also the lead has far to high resistance for the stated current.

But the worst bit for me was the lack of separation between the mains winding and the low-voltage winding. Having just a plastic tape as a barrier that is almost guaranteed to fail when the transformer gets hot is likely to be fatal!

Reminds me of a few years back, when there were brake shoes on sale in the UK (from China) which would fail on an emergency stop. And Electric isolators (same source) which wouldn't isolate in an emergency.

Frightening.

shaurz
Posts: 49
Joined: Thu Nov 03, 2011 12:09 pm

Re: Power supply WARNING

Mon Jan 09, 2012 1:09 pm

I've bought a few supposedly switching power supplies from eBay. Guess I'll have to do some testing to see if they're OK or should be thrown away.

BlueClogger
Posts: 34
Joined: Mon Dec 26, 2011 11:01 am

Re: Power supply WARNING

Mon Jan 09, 2012 2:21 pm

All the clever chatter above is pointless - the heart of the matter is that some PSUs are not what they state, and may be dangerous to life and Pi.

After reading this topic I've knocked together a simple PSU tester using a USB extension lead split into a terminal block.  The shielding and data lines pass straight through.  The power lines have loops that allow a multimeter in series with the +ve (10A) and another in parallel (20V) to have their probes screwed into the block terminals.  I connected a four-way hub to the output and tested my HP iPaq PSU (made in China) rated at 5V 1A.  I connected up to 3 USB 2.5" drives to the hub and got the following values: open – 5.12V, 0A; single drive – 4.99V, 0.32A; two drives – 4.89V, 0.65A; three drives – 4.78V, 0.97A.

I think this will be my first choice for the Raspi, with a powered hub on the USB output.

I took a picture, but I have no web starage, and can't get it to display from my HD. Anyone know how?

Prometheus
Posts: 308
Joined: Tue Dec 13, 2011 11:09 pm

Re: Power supply WARNING

Mon Jan 09, 2012 2:58 pm

BlueClogger said:

I took a picture, but I have no web starage, and can't get it to display from my HD. Anyone know how?
Go to imgur, upload the image, and then once it's uploaded, copy the code that they give you that's labelled "BBCode (message boards & forums)", and paste it into a post.

That ought to do it.

BlueClogger
Posts: 34
Joined: Mon Dec 26, 2011 11:01 am

Re: Power supply WARNING

Mon Jan 09, 2012 7:15 pm

Many thanks, Prometheus.

Of course, you do need the two multimeters (they're just £5 jobs from Maplin) including one that will measure currents in excess of the supply!

With the connections numbered 1-12 (odd numbers at the top) the braid goes into 7 and out of 8. The +ve goes into 3 and is looped to 1. It comes out at 6. The negative goes in at 9 and out at 10, and is looped to 11. 4 (+) and 5(-) are the ammeter connections. This could be replaced by a plain link if you are not measuring current. 2 & 12 are the volmeter connections.

Doimg it this way means that the meters can be connected easily without interfering with the (rather thin) USB wires.

The readings are showing 5.02V and 036A with a single drive attached to the output.


jwatte
Posts: 203
Joined: Sat Aug 13, 2011 7:28 pm

Re: Power supply WARNING

Mon Jan 09, 2012 11:42 pm

9V DC power supply I found a while back:

https://plus.google.com/115565509658927 ... d1obRgDLjn

I find that the power supplies for the Nintendo Gameboy DS series to be very well built, though. The wrong plug, but that's easily fixable

jwatte
Posts: 203
Joined: Sat Aug 13, 2011 7:28 pm

Re: Power supply WARNING

Mon Jan 09, 2012 11:55 pm

Burngate said:


Just to obfuscate: Ohms law doesn't always apply – think of a zener!


Ohm's law holds just fine! A Zener diode is a voltage-dependent resistor. There are other kinds of dependent resistors, including heat-dependent (PTC, NTC), control-current-dependent (BJ transistors), control-voltage-dependent (FE transistors), light-dependent (phototransistors), frequency-dependent (capacitors) etc.

The Zener diode can be seen as varying its resistance based on the voltage across it and current through it so that it keeps the rated voltage across it. This is why it can't "clamp" the voltage in parallel across it to a value you want (like a regulator) on its own -- it will just create a short, and burn out.

User avatar
Burngate
Posts: 6059
Joined: Thu Sep 29, 2011 4:34 pm
Location: Berkshire UK Tralfamadore
Contact: Website

Re: Power supply WARNING

Tue Jan 10, 2012 1:17 pm

Ohm's Law: V=IR, or R=V/I
The implication is that R is a constant for a given object. Turns out to be not neccessarily true for all objects.
So we redefine Ohm's Law: r=dV/dI
r is now a variable, and is calculable. Ohm's Law is now not a law, just a definition of resistance, which can change with circumstances.
Power supplies: for simplicity, we assume it can be modelled as a constant voltage source with a built-in (constant) source impedance.
Then measuring its open-circuit (no load) voltage gives us one parameter, and measuring its short-circuit current gives us the other.
But if we have something - a length of naff cable for example - between the PSU and our meters, we are going to get a different (wrong?) answer. Having another meter gives us the opportunity to characterise that naff cable as well.
I buy a PSU complete with cable, and it says "5V 1A", I expect to get that out of the end of the cable. If I find I can't, I can look into it. Maybe it's just the naff cable, maybe the PSU before the cable, or maybe both. Interesting, either way, but at the end of the day, not useful - I can't ask the supplier for just a new cable or just a new PSU, he'll just replace them as one item.
More to the point, even if that PSU had done what the tin said - give out 1A @ 5V - by the look of it it could roast my Pi, kill me, or burn my house down. Or someone else. Not only should we keep well away from these things, we should make a song and dance to get them taken off the market.
Like nuke China, perhaps

Wooloomooloo
Posts: 92
Joined: Fri Nov 25, 2011 10:52 am

Re: Power supply WARNING

Tue Jan 10, 2012 1:40 pm

jwatte said:


Burngate said:


Just to obfuscate: Ohms law doesn"t always apply – think of a zener!


Ohm"s law holds just fine! A Zener diode is a voltage-dependent resistor. There are other kinds of dependent resistors, including heat-dependent (PTC, NTC), control-current-dependent (BJ transistors), control-voltage-dependent (FE transistors), light-dependent (phototransistors), frequency-dependent (capacitors) etc.

The Zener diode can be seen as varying its resistance based on the voltage across it and current through it so that it keeps the rated voltage across it. This is why it can't "clamp" the voltage in parallel across it to a value you want (like a regulator) on its own -- it will just create a short, and burn out.



LOL, right. In fact, capacitors are magic resistors! One can charge them up, disconnect them, and have nonzero voltage at zero current! Weeeee!

Also in other news, quantum physics is merely Newtonian laws with some varying constants...

Return to “General discussion”